A Case of Chip Multithreading Architecture with Resource Unit Manager

Juan Fang, Zhenxi Pan, Lu Yu, Sitong Liu
{"title":"A Case of Chip Multithreading Architecture with Resource Unit Manager","authors":"Juan Fang, Zhenxi Pan, Lu Yu, Sitong Liu","doi":"10.1109/ISCC-C.2013.8","DOIUrl":null,"url":null,"abstract":"In order to improve the overall throughput of processors, to increase thread-level parallelism (TLP) and instruction-level parallelism (ILP) intra and inter cores, and ensure the scalability of multi-core systems, can a novel mechanism chip multithreading with resource unit manager (RUM-CMT) be used, which is presented in this paper. By introducing the efficiency difference of resource utilization among threads, such as, the condition of dynamically resource allocation to CMT architecture and add an absorption barrier for all threads to make their reserved resources far away from hidden hunger. For verify the applicability of RUM-CMT, a 4 cores with 2 threads processor and a 4 cores with 4 threads processor are simulated based on GEMS & Simics simulator. And this limited study shows RUM-CMT can increase performance by 29.79%, and an improve success committed instructions per squash instructions by 31.15% and fairness attains 6.25% improvements over static partition on 4 cores with 4 threads processor. The 4 cores with 2 threads also attains passable enhance. So RUM-CMT can accurately and concise reallocate resource according to phase turning in the running of applications, improve resource utilization, mining the ILP and TLP in applications. In other words, RUM-CMT improves the controllability of threads, reduces invalid losses and improves system efficiency, and it is a suitable solution for multicore, multithreaded architecture.","PeriodicalId":313511,"journal":{"name":"2013 International Conference on Information Science and Cloud Computing Companion","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2013-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2013 International Conference on Information Science and Cloud Computing Companion","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ISCC-C.2013.8","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

Abstract

In order to improve the overall throughput of processors, to increase thread-level parallelism (TLP) and instruction-level parallelism (ILP) intra and inter cores, and ensure the scalability of multi-core systems, can a novel mechanism chip multithreading with resource unit manager (RUM-CMT) be used, which is presented in this paper. By introducing the efficiency difference of resource utilization among threads, such as, the condition of dynamically resource allocation to CMT architecture and add an absorption barrier for all threads to make their reserved resources far away from hidden hunger. For verify the applicability of RUM-CMT, a 4 cores with 2 threads processor and a 4 cores with 4 threads processor are simulated based on GEMS & Simics simulator. And this limited study shows RUM-CMT can increase performance by 29.79%, and an improve success committed instructions per squash instructions by 31.15% and fairness attains 6.25% improvements over static partition on 4 cores with 4 threads processor. The 4 cores with 2 threads also attains passable enhance. So RUM-CMT can accurately and concise reallocate resource according to phase turning in the running of applications, improve resource utilization, mining the ILP and TLP in applications. In other words, RUM-CMT improves the controllability of threads, reduces invalid losses and improves system efficiency, and it is a suitable solution for multicore, multithreaded architecture.
基于资源单元管理器的芯片多线程架构实例
为了提高处理器的整体吞吐量,提高内核内和内核间的线程级并行性(TLP)和指令级并行性(ILP),保证多核系统的可扩展性,本文提出了一种基于资源单元管理器(RUM-CMT)的芯片多线程机制。通过在CMT体系结构中引入线程间资源利用效率差异,如动态资源分配条件,并为所有线程添加吸收屏障,使其保留的资源远离隐性饥饿。为了验证RUM-CMT的适用性,在GEMS和Simics模拟器上分别对4核2线程处理器和4核4线程处理器进行了仿真。这项有限的研究表明,与4核4线程处理器上的静态分区相比,ram - cmt可以将性能提高29.79%,将每个压缩指令的成功率提高31.15%,公平性提高6.25%。4核2线程也获得了尚可的增强。因此,RUM-CMT可以在应用程序运行过程中根据阶段转换准确、简洁地重新分配资源,提高资源利用率,挖掘应用程序中的ILP和TLP。也就是说,ram - cmt提高了线程的可控性,减少了无效损耗,提高了系统效率,是一种适合多核、多线程架构的解决方案。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信